Resolve uncertainty early.
Through focused technical workshops, I help your team clarify requirements, expose constraints, and test assumptions before they become expensive implementation changes.
Independent engineer & technical partner
I'm Adrian Topala, a hands-on senior engineer and tech lead. I help teams build AI-enabled applications, modernise critical platforms, and deliver reliable software—with clear priorities and fewer costly detours.
Trusted by
Through focused technical workshops, I help your team clarify requirements, expose constraints, and test assumptions before they become expensive implementation changes.
I draw on 20+ years across projects, teams, and industries to recognise likely pitfalls and narrow the options. Decisions and tradeoffs stay transparent, so everyone understands the approach and why it fits.
I combine hands-on engineering with AI-assisted implementation and testing to shorten feedback cycles. Each iteration should resolve important questions and move the product closer to release.
A shorter path to market: resolve uncertainty early, make sound decisions, and reduce avoidable iterations and rework.
01 / Selected work
The challenges I took on, the decisions I owned, and the results for the business.
A complex request-for-quotation workflow needed modernising. I started with a lean implementation of that critical workflow, then re-architected the frontend for modular delivery.
Role: Lead Software Engineer / Tech Lead Frontend at Techpilot / DynamicMarkets GmbH.
I started with a lean version of the platform's most complex workflow, introduced a micro-frontend architecture, and worked with CTO and COO stakeholders to align the technical direction with business priorities.
Result: RFQ processing time reduced by a factor of 20.
TypeScript · React · Redux · Micro-frontends
Taking a computer-vision checkout pilot into production required coordinated delivery across software, devices, and external teams. I led engineering and established the integration and release approach.
Role: Lead Software Engineer / Tech Lead for a major European grocery retail group, via Reply Germany.
I architected the platform, built a virtual-store simulator, and established integration contracts, release cadences, and rollout readiness criteria across teams.
Result: A rollout from a single-store pilot to multiple stores and countries. The solution received the RETA 2023 award for Best AI & Robotics.
Computer vision · Node.js · React · Azure · Kubernetes
Complex Excel-based reporting made cross-division risk data difficult to manage. I replaced those workflows with version-controlled dashboards and automated data integration.
Role: Senior Consultant at Siemens AG, via Pentland Firth Software GmbH.
I led the data engineering and BI transformation, automated migration of Excel logic into QlikView, and introduced parallel delivery pipelines for the programme.
Result: Sixfold improvement in data integration speed, supporting cross-division risk analytics.
QlikView · Data engineering · ETL · Risk analytics
02 / AI in practice
Founder & principal engineer
A pre-release platform that turns database schemas and business context into AI-native data applications. Operational data and query execution stay in the customer's environment.
I also maintain the Apache-2.0 data framework behind this work. It gives applications and AI services a type-safe, governed route to operational data. Its MCP integration lets agents discover approved schemas and compose bounded read-only queries, with support for PostgreSQL, SQL Server, SQLite, and DuckDB.
03 / Working together
AI integration, hands-on implementation, and technical leadership to move your product forward.
I connect AI services to your applications and business data: assistants, retrieval-augmented search, and workflow automation, with explicit permissions and validation.
A good fit: a team turning an AI prototype into a dependable feature, or adding AI to an existing application.
Discuss AI integrationI join your team to own technical decisions, align the people involved, and move delivery forward—from architecture through implementation and rollout.
A good fit: a complex programme that needs an experienced, hands-on technical lead.
Discuss technical leadershipI take ownership of difficult engineering work: modernising an existing platform, building a critical feature, or connecting applications, data, and infrastructure.
A good fit: a team that needs senior delivery capacity and clear communication.
Discuss engineering supportI help define your technical direction, shape the delivery model, and support engineering decisions at the leadership table—with a practical path to handover.
A good fit: a startup or growing company that needs technical leadership without a full-time CTO.
Discuss technical directionI use AI to support implementation, investigation, and test generation. Clear specifications, human code review, and automated validation keep the work grounded in your requirements. The aim is faster progress with software your team can understand and maintain.
04 / About Adrian
I've worked across startups and large enterprises, building web, cloud, mobile, IoT, and desktop systems. My work combines architecture and technical leadership with writing and shipping production code.
That has included scaling an R&D centre, leading cross-functional programmes, and building platforms from the first implementation through rollout. Today, I also build Vexnor, an AI-native data application platform and open-source framework.
I bring a practical perspective to technical decisions: understand the problem, make the tradeoffs clear, and keep the team moving towards a working result.
View my LinkedIn profile ↗Building a pre-release, AI-native data application platform that turns database schemas and business context into customer-controlled applications with governed metrics, dashboards, APIs, collaboration, and AI-assisted exploration. Designing and implementing the platform hands-on across TypeScript, React, Node.js, PostgreSQL, DuckDB, Docker, and cloud infrastructure.
Designed the platform so operational data and query execution remain in the customer's environment. Working with early customer design partners to validate real workflows, deployment requirements, and product direction.
Created and maintain an Apache-2.0, type-safe data framework designed to help existing applications integrate AI services with operational data through a governed interface. It combines schema-derived TypeScript models, typed CRUD, and real SQL with compile-time inference of results and required parameters.
Built isomorphic browser/server query execution without a separate endpoint for every query, portable contracts across TypeScript, .NET, and Go, and support for PostgreSQL, SQL Server, SQLite, and DuckDB. Its local MCP server lets AI agents discover approved schemas, resolve foreign-key paths, and compose bounded read-only queries without emitting SQL.
Built NIFAM, an interactive web UI for the decisionSupport R package (FastAPI · Vue 3 · TypeScript · R/Plumber), using LLMs for model explanation, specification clarification, and user-facing guidance via RAG while retaining deterministic execution.
Prototyped a configurable IoT self-checkout system spanning kiosk client, backend services, and OTA pipelines, and implemented an unpublished real-time tactical puzzle game in PhaserJS with procedurally generated levels.
Architected and led delivery of a frictionless, computer-vision-based IoT checkout platform, initially piloted in a single store and progressively scaled to multiple stores and countries. Delivered an award-winning AI-powered retail solution (RETA Award 2023 — Best AI & Robotics), reducing checkout time by 90%+.
Built a production-grade virtual store and customer-journey simulator enabling any team member or executive to simulate real customer behaviour, providing continuous visibility into system maturity. Orchestrated cross-functional delivery across backend, frontend, mobile, IoT, UX, BI, QA, and external vendors by establishing clear integration contracts, release cadences, and rollout readiness criteria.
Stabilised HelloCoins.com crypto exchange with a 6-engineer remote team. Set up Node.js · React · MongoDB stack, CI/CD pipelines, and agile delivery. Facilitated collaboration across engineering functions and aligned delivery milestones with product and business expectations.
Initiated modernisation by rebuilding the core RFQ workflow as a lean MVP — the system's most complex and business-critical feature — after earlier freelance attempts failed. Re-architected Europe's largest manufacturing sourcing platform with micro-frontends (TypeScript, React, Redux, Material-UI) for better modularity and team autonomy.
Reduced RFQ processing time by 20×. Collaborated closely with CTO/COO-level stakeholders, aligning technical direction with business priorities and ensuring consistent delivery across engineering and product teams.
Architected a private pricing platform turning brand communities into micro-influencers. Integrated Shopify API and headless CMS to automate product onboarding, pricing sync, and order management.
Delivered autoscaling live/VOD streaming platforms on GCP with Node.js micro-services and Wowza servers for Limelight Music (US) and Digita Oy (Finland).
Led design and development of a fraud-detection and customer-risk system for IoT self-scanning carts. Coordinated a cross-functional Agile team across business, UX, and engineering.
Raised review rate for active customers from 18% to 94% in under 3 months. Standardised release and deployment management and implemented an OTA (over-the-air) IoT deployment system. Delivered on Azure · Node.js · React · .NET Core · C# · MongoDB · SQL Server.
Led multi-division data engineering and BI transformation for enterprise risk reporting, migrating complex Excel-based workflows into scalable, version-controlled QlikView dashboards.
Automated Excel-to-QlikView logic migration, improving data integration speed by 6× and enabling cross-division analytics. Implemented parallel Kanban pipelines and DevOps processes to enhance efficiency in a complex, multi-division environment.
Designed and built an IoT solution integrating commodity GSM/Android thermal printers with a centralised ERP for a chain of cafes in Melbourne — connecting order taking, kitchen, and delivery operations end-to-end. Sole architect, developer, and delivery owner.
Architected and developed a consumer-facing SaaS extension for the Dracoon secure file-sharing ecosystem, enabling simplified document sharing via dynamically generated QR codes.
Co-founded a stealth-mode security start-up in Munich.
Architected and developed the high-performance .NET/C#/ASP.NET API layer that integrated iPad and Windows desktop clients with Brainloop's core services, driving a 30% increase in customer acquisition upon market release. Delivered Client SDKs (.NET, Java, iOS/Xamarin) with libraries, code samples, and documentation.
Built a chatbot for automating sales conversions of an online marketing business. Sole architect, developer, and delivery owner.
Directed technical delivery and architecture of enterprise-scale nearshoring projects for major Dutch clients (Hyundai, Group M, Boskalis) using .NET/C#/ASP.NET.
Managed project budgets up to €2M and led cross-functional teams of 10+ engineers. Core team member that founded the Iasi R&D centre, growing it from 4 to 30 engineers.
For Boskalis, architected and led development of SOPRO — a complex SOA web system for managing soil processing activities worldwide, integrating SCADA, BizTalk orchestration, Active Directory, and SQL Server Reporting Services. Also delivered SOFLO — a soil flow registration and dynamic reporting system for Dutch government compliance. Both projects followed PRINCE2 with iterative Scrum, TDD, and CI on Team Foundation Server.
Built an eCommerce solution for a handmade jewellery start-up. Sole architect, developer, and delivery owner.
Built a classroom management system for an education services start-up. Sole architect, developer, and delivery owner.
Lead developer role delivering software solutions for local and international clients.
Built a VoIP PIN sales management solution for a small telecom services provider. Sole architect, developer, and delivery owner.
Early career software developer role.
Full-spectrum technical expertise across modern engineering disciplines
Designing AI-enabled products and delivery workflows with explicit guardrails: AI service integration, schema-guided tools, MCP servers, RAG, spec-driven development, AI-supported test criteria, and deterministic validation — with human review and production-grade engineering throughout.
Microsoft Imagine Cup 2005 — 1st Place, Web Development
Yokohama, Japan · Built a platform integrating AI, machine learning, neural networks, and fuzzy logic to bridge medical and IT fields.
Available for select freelance engagements
In our first conversation, we'll establish whether I'm the right fit and identify a sensible starting scope.
Prefer another way? Open the form in a new tab ↗ or contact me on LinkedIn ↗.